Data and analytics have dramatically changed the advertising industry’s ability to assess and understand human behavior, allowing marketers to better reach and speak to their audiences. Despite its negative press coverage in the past months, data can be used for much more than just selling you that new pair of shoes (I don’t know who needs to hear this today, but you don’t need that new pair of shoes!). Data-backed communications can be used strategically in creating change across some of the most pressing issues of today.
How can we use data for social good?
Data is a very broad term, especially in the context of advertising. In the most straight-forward sense of the term, data can help with storytelling and provide perspective to an issue. For someone who has never experienced hunger, the magnitude of the issue of food insecurity in the US may not be apparent. For someone who doesn’t come from an area where it is common to own guns, family fire may not seem like a common or relevant issue, and an advertiser’s perception of a gun owner may be severely misinformed. We cannot tackle these social issues without being able to communicate their importance - not just from a moral standpoint, but from the context of proximity and scale.
Sifting through large amounts of data to convey a powerful yet digestible story and using data to paint a fuller picture of these issues, when paired with information on actions the audience can take to make a difference, is where the true power of analytics lies. Data can help draw attention and help validate a perspective in objectivity, as well as give insights into an issue through patterns that may have otherwise been overlooked.
Artificial Intelligence (AI) and Machine Learning (ML) are use cases of data that have been both the most contended and on the rise in recent years. Google “AI scandal” and you’ll get hundreds of articles about algorithms gone wrong. However, data and machine learning can also do a lot of good. They can help identify users more interested in buying sustainable clothing, build issue awareness for a cash-strapped non-profit, or reach users more likely to be unaware, and in need, of community resources in their area. In fact, much of the most recent research in Machine Learning and AI is focused on larger societal questions – including searching for early warning signs of opioid addiction, investigating pathways out of poverty, recognizing and flagging hate speech online, and offering medical professionals search data to help identify infectious disease outbreaks.
Building Ethical Data
It is also important to think of data at its current scale in the context of data collection methodology and ethical design, not just in the context of the data itself. How can we build systems and algorithms that mitigate the human bias of the developers who build them? How can we create digital products with accessibility in mind? How can we ensure data collection is ethical and does not infringe on the rights or the security of the user?
With the amount of data and data sources that are available, complete oversight is nearly impossible. Furthermore, most data-based tools in advertising are relatively new. We therefore need to think of consumers as partners in this process rather than treating them as the bottom line.
Accountability and Accessibility of Information
Despite most algorithms being the proprietary information of the tech companies that builds them, having accountability for the effects, intentional or not, of an algorithm or program is essential. Transparency, power, and intent are the difference between persuasion and manipulation. Therefore, as a user I should be able to access information, in layman’s terms, on why I am seeing an ad. What data is being collected from me? How is my data being collected and categorized? Do I find this categorization represents me accurately? Allowing for this conversation and building a feedback loop opens opportunities for the user-advertiser relationship to become more mutually beneficial.
Part of this collaboration in machine building includes representation of user perspectives. The tech industry has been notoriously homogenous, and while efforts to increase diversity and inclusion within the industry have begun, there is still a long way to go. The lack of diversity spans recruiting practices, such as reliance on referrals, segregation within early and university level education systems, and lack of representation and welcoming of minorities within the industries. The tech and ad industries need to work together to create systems of opportunities for minorities to become involved and trained in these industries, and then actually be hired and heard within these jobs. Diverse experiences are important from a moral, economic and innovation standpoint for companies and employees alike.
Certain best practices are already in place when it comes to working with data, such as asking for explicit permission from the user to repost user-generated content, avoiding collecting personally identifiable information (PII), or asking users to explicitly opt into data collection when entering a site. However, even these practices are not always user friendly, and assume a certain level of understanding from the user of how data collection and algorithms work.
Because of the relationship between the advertising industry and technology companies, who often rely on advertising, it is part of our responsibility as advertisers to advocate for more transparency within data, and ethical data governance, collection, and usage. As information availability expands, we must keep these guiding principles in mind to promote data usage in a meaningful, purpose-driven way.